<p>Electrical fault classification is vital for ensuring the reliability and safety of power systems. Accurate and efficient fault classification methods are essential for timely and effective maintenance. In this paper, we propose a novel approach for effective fault classification through Grassmann manifolds, a non-Euclidean space that captures the intrinsic structure of high-dimensional data and offers a robust framework for feature extraction. We use simulated data for electrical distribution systems with various types of electrical faults, including systems integrated with distributed generators. The proposed method involves transforming the raw measurement fault data into Grassmann manifold space using an autoregressive moving average (ARMA) model and techniques from differential geometry. This transformation aids in uncovering the underlying fault patterns and reducing the computational complexity of subsequent classification steps. To achieve fault classification, we employ a support vector machine (SVM) classifier with a Gaussian kernel, optimized to operate within the Grassmann manifold space. This enables effective discrimination between different fault classes, including no-fault conditions. We adopt a comprehensive 10-fold cross-validation strategy to ensure robust model performance assessment and generalization capabilities. The results demonstrate the superior performance of our proposed approach, achieving the highest accuracy (99.96%) among compared models, outperforming both conventional machine learning and state-of-the-art deep learning techniques. The proposed method offers significant advantages in terms of computational efficiency, ease of implementation, and reduced memory requirements, while maintaining or surpassing the accuracy of traditional neural network models. This innovative approach showcases its ability to accurately differentiate between various fault types in electrical distribution systems, potentially revolutionizing fault classification in modern power systems.</p>